‘Spooky’ Signal Still Haunts Paclitaxel DCBs in PAD, but Updated Analysis Reassures

SAN FRANCISCO, CA—In separate presentations here, experts continued to struggle with what one moderator called the “spooky” signal of mortality for paclitaxel-coated devices used to treated PAD. Packed presentations yesterday at TCT 2019 included, on the one hand, reassuring results from an updated Medicare analysis suggesting that survival is no different in real-world patients treated with drug-coated versus non-drug-coated devices. But, on the other hand, Konstantinos Katsanos, MD, PhD (Patras University Hospital, Rion, Greece), the lead author of the meta-analysis that sent the endovascular community into a tailspin late last year, was back with new data hinting at more potential harm from paclitaxel.

Mitchell W. Krucoff, MD (Duke University Medical Center, Durham, NC), who moderated an FDA Town Hall session earlier this week expressed both ongoing frustration and scientific curiosity over how to explain the mortality signal. That signal came to light in the Katsanos paper published late last year in the Journal of the American Heart Association. Since then, the US Food and Drug Administration has issued three communications on the topic, most recently an updated letter in August, summarizing what they know about the signal and to recommend that clinicians closely monitor patients implanted with the devices and obtain informed consent before any paclitaxel-based device procedures.

According to Krucoff, responding to multiple presentations reviewing what is known and still unknown, said the distribution of the trajectory of late mortality across such different devices and randomized trials, “is spooky and hard to explain, even as we all sort of tear our hair out trying to understand what is the biological mechanism.”

Staying with the supernatural theme, the co-moderator of another TCT session Friday was prompted to ask, “Are we hunting a ghost?” That was in response to a presentation by Eric A. Secemsky, MD (Beth Israel Deaconess Medical Center, Boston, MA), updating his previously published meta-analysis. His latest report showed no difference in survival among more than 150,000 Medicare beneficiaries who have been followed out to a median of just over 2 years, with longest follow up of 4 years.

We’re not there yet to completely exonerate what was found in the [Katsanos] meta-analysis, but I think we’re moving closer to saying at least that the use of these devices and the guidance from the FDA to not remove the devices from the market is reasonable because we have so much data that cannot replicate the signal of harm. Eric Secemsky

These data add strength to two prior Medicare analyses, and for the first time include a low-risk population of over 15,000 patients under age 70 with fewer than two comorbidities and no critical limb ischemia. Secemsky showed that the cumulative incidence of death in these patients was 19.9% in the non-drug-coated group and 18.0% in the paclitaxel-coated group (adjusted HR 0.87; 95% CI 0.79-0.95).

Secemsky also announced that the Medicare cohort will be part of a new prospective study done in conjunction with the FDA called SAFE PAD CMS, “which is looking at this entire population as it moves to the 5-year median follow-up period,” he noted.

In an interview with TCTMD, Secemsky said that, in an attempt to raise the study to the standards of a randomized trial, it will include dummy variables and confounders “to make sure we see the same signal of safety” while also attempting to answer additional questions that could not be addressed by randomized trial data.

“We know there are limitations of real-world data, which I respect, but for us to be able to look at these data in multiple perspectives, multiple groups, different approaches statistically and see evidence of no harm associated with these devices is a very reassuring step to saying maybe this is only a signal and not a true causative relationship,” Secemsky commented. “We’re not there yet to completely exonerate what was found in the [Katsanos] meta-analysis, but I think we’re moving closer to saying at least that the use of these devices and the guidance from the FDA to not remove the devices from the market is reasonable because we have so much data that cannot replicate the signal of harm.”

In the original meta-analysis of 28 trials led by Katsanos, the risk ratio for all-cause death at 5 years was 1.93 (95% CI 1.27-2.93) for paclitaxel versus uncoated devices. Updated data presented by Katsanos here puts the increased risk of death at 5 years at 62% for the paclitaxel devices compared with uncoated devices, but with a 42% decrease in TLR. In terms of absolute risk, this translates to 13% fewer TLRs at the expense of 6% more deaths with paclitaxel, Katsanos noted.

“So, one out of eight will benefit from avoiding one TLR, one out of 16 may die from the increased risk of all cause mortality, in very, very simple terms, from the randomized studies to 5 years,” he said. Katsanos reiterated that most trial sponsors have refused to give him access to patient-level data for further analysis, so his group has continued to update their work with summary-level data.

Amputations and QoL Questions

The update from Katsanos also provided new numbers on leg amputation, suggesting a higher rate with paclitaxel. Only eight trials have reported that event, with an absolute risk of 1.0% for paclitaxel versus 0.3% for non-drug-coated device.

“Beware those are very, very rare events so that signal is unstable by definition,” Katsanos told TCTMD.

While many clinicians maintain that paclitaxel-based devices improve quality of life (QoL), here too there is little in the way of data to support that claim. Katsanos said only ILLUMENATE, IN.PACT JAPAN, IN.PACT SFA, and RANGER have collected QoL data out to 1 year.

“Basically, the pooled estimate shows no significant [QoL] benefit whatsoever,” Katsanos said. “There is a [suggestion] that the collection of the quality of life data was not performed at regular time intervals to be able to pick up any signals to correlate with the TLR benefit. Basically, the trials weren't designed to do this collection properly.” He added that while there is no question that the paclitaxel-based devices are effective in reducing TLR at 1 year and beyond, “the magnitude of the benefit, regardless of statistical significance, doesn’t seem to imply that it is large enough to justify the associated risk.”